一种面向多星多分辨率的SAR图像舰船候选区域提取方法

胡炎; 单子力; 高峰

doi:10.11999/JEIT180525

一种面向多星多分辨率的SAR图像舰船候选区域提取方法

doi: 10.11999/JEIT180525

中国电子科技集团公司航天信息应用技术重点实验室石家庄 050081

基金项目: 中国电子科技集团公司航天信息应用技术重点实验室开放基金(EX166290025)

详细信息

作者简介:
胡炎：男，1991年生，硕士，工程师，研究方向为图像智能处理技术与机器学习

单子力：男，1980年生，博士，高级工程师，研究方向为微波遥感与大系统集成技术

高峰：男，1978年生，学士，高级工程师，研究方向为航天应用总体设计

通讯作者:
单子力　huyantju@126.com

中图分类号: TN957.52
计量
- 文章访问数: 2369
- HTML全文浏览量: 618
- PDF下载量: 122
- 被引次数: 0
出版历程
- 收稿日期: 2018-05-29
- 修回日期: 2018-12-18
- 网络出版日期: 2018-12-26
- 刊出日期: 2019-04-01

Candidate Region Extraction Method for Multi-satellite and Multi-resolution SAR Ships

CETC Key Laboratory of Aerospace Information Applications, Shijiazhuang 050081, China

Funds: The Open Foundation of CETC Key Laboratory of Aerospace Information Applications (EX166290025)

摘要

摘要:
基于CFAR和核密度估计(KDE)的SAR传统舰船候选区域提取方法存在以下缺陷：CFAR虚警率依赖人工经验选择；CFAR仅对杂波分布建模，会对被检目标构成一定的漏检风险；利用KDE进行强海杂波过滤时，需凭人工经验选择滤除阈值。这使得传统舰船候选区域提取方法无法适应多星多分辨率等复杂场景。该文提出一种面向多星多分辨率的SAR图像舰船候选区域提取算法，针对CFAR算法的缺陷，提出采用均值二分法迭代逼近目标计算分割阈值，在克服CFAR缺陷的同时，计算效率比CFAR提高10倍以上；针对KDE的缺陷，提出了区块KDE结合大阈值滤除强海杂波，再借助种子点生长算法重建目标。由于大阈值具有足够的阈量，使得算法可以适应更复杂的场景。实验表明所提方法具有不漏检、阈值自适应、计算效率高、虚警率低的优点，具备优秀的多星多分辨率SAR舰船候选区域提取能力。
- 图像处理 /
- 舰船候选区域 /
- 均值二分法 /
- 目标重建 /
- 种子点生长 /
- 阈值自适应
Abstract:
The traditional methods based on CFAR and Kernel Density Estimation (KDE) for SAR ship candidate region extraction has the following defects: The choice of false alarm rate of CFAR depends on artificial experience; CFAR only models the sea clutter distribution, which poses a certain risk of missing detection to the target; When KDE is used to filter strong sea clutter, the threshold must be selected by artificial experience. These defects make the traditional method unable to adapt to complex scene, such as multi-satellite and multi-resolution. A candidate region extraction method for multi-satellite and multi-resolution SAR ships is proposed. In view of the defects of CFAR, an iterative method of mean dichotomy is proposed to approximate the target and calculate the segmentation threshold. The calculation efficiency of this method is more than 10 times higher than that of CFAR while overcoming the defects of CFAR; In view of the defects of KDE, block KDE combined with large threshold is used to filter strong sea clutter, and then seed point growth algorithm is used to reconstruct target. Because the large threshold has enough thresholds, the method can adapt to more complex scenarios. Experiments show that the proposed method has the advantages of no missed detection, self-adaptive threshold, high computational efficiency, and low false alarm rate. It has excellent multi-satellite and multi-resolution SAR ship candidate region extraction capability.
- Image processing /
- Ship candidate region /
- Mean dichotomy /
- Target reconstruction /
- Seed point growth /
- Threshold adaptive

HTML全文

图 1 目标候选区域提取算法流程

下载: 全尺寸图片幻灯片

图 2 阈值变化曲线示意图

下载: 全尺寸图片幻灯片

图 3 迭代次数分别为1, 5, 10, 20, 100的实验结果

下载: 全尺寸图片幻灯片

图 4 G分块与舰船的3种交叠关系

下载: 全尺寸图片幻灯片

图 5 BKDE结合大阈值滤除强海杂波实验结果

下载: 全尺寸图片幻灯片

图 6 种子点生长算法邻域模板

下载: 全尺寸图片幻灯片

图 7 多场景舰船候选区域提取实验结果

下载: 全尺寸图片幻灯片

图 8 舰船候选区域精准提取实验结果

下载: 全尺寸图片幻灯片

图 9 本文算法与瑞利CFAR实验对比

下载: 全尺寸图片幻灯片

图 10 本文算法与瑞利-CFAR在复杂海况下的检测结果对比

下载: 全尺寸图片幻灯片

图 11 算法运行时间

下载: 全尺寸图片幻灯片

表 1 典型舰船尺度表

船舶类型	船舶名称	船长(m)	船宽(m)	船舶类型	船舶名称	船长(m)	船宽(m)
集装箱船	COSCO_KAWASAKI	260	32	油船	ZHONG CHI	188	31
	N.Y.K.LEO	300	40	油船	WANG CHI	187	32
	HYUNDAI_BRIDGE	182	35	驱逐舰	055型	160～180	21～23
	HUA_RUN_CHUANG_YE	190	31	驱逐舰	日本金刚级	161	21

下载: 导出CSV

表 2 该实验场景下6种分布CFAR的P_fa的合理取值表

分布名称	P_fa (%)	分布名称	P_fa (%)
高斯分布	2～3	指数分布	10～15
韦布尔分布	3～4	Gamma分布	2～4
对数分布	5～10	瑞利分布	2～4

下载: 导出CSV

参考文献(16)

陈琪, 王娜, 陆军, 等. SAR图像港口区域舰船检测新方法[J]. 电子与信息学报, 2011, 33(9): 2132–2137 doi: 10.3724/SP.J.1146.2011.00018

CHEN Qi, WANG Na, LU Jun, et al. A new method for ship detection in harbor region of SAR images[J]. Journal of Electronics &Information Technology, 2011, 33(9): 2132–2137 doi: 10.3724/SP.J.1146.2011.00018

文伟, 曹雪菲, 张学峰. 一种基于多极化散射机理的极化SAR图像舰船目标检测方法[J]. 电子与信息学报, 2017, 39(1): 103–109 doi: 10.11999/JEIT160204

WEN Wei, CAO Xuefei, and ZHANG Xuefeng. PolSAR ship detection method based on multiple polarimetric scattering mechanisms[J]. Journal of Electronics &Information Technology, 2017, 39(1): 103–109 doi: 10.11999/JEIT160204

艾加秋, 齐向阳, 禹卫东. 改进的SAR图像双参数CFAR舰船检测算法[J]. 电子与信息学报, 2009, 31(12): 2881–2885 doi: 10.3724/SP.J.1146.2008.01707

AI Jiaqiu, QI Xiangyang, and YU Weidong. Improved two parameter CFAR ship detection algorithm in SAR images[J]. Journal of Electronics &Information Technology, 2009, 31(12): 2881–2885 doi: 10.3724/SP.J.1146.2008.01707

胡炎, 单子力, 高峰. 基于Faster-RCNN和多分辨率SAR的海上舰船目标检测[J]. 无线电工程, 2018, 48(2): 96–100 doi: 10.3969/j.issn.1003-3106.2018.02.04

HU Yan, SHAN Zili, and GAO Feng. Ship detection based on faster-RCNN and multi-resolution SAR[J]. Radio Engineering, 2018, 48(2): 96–100 doi: 10.3969/j.issn.1003-3106.2018.02.04

徐丰, 王海鹏, 金亚秋. 深度学习在SAR目标识别与地物分类中的应用[J]. 雷达学报, 2017, 6(2): 136–148 doi: 10.12000/JR16130

XU Feng, WANG Haipeng, and JIN Yaqiu. Deep learning as applied in SAR target recognition and terrain classification[J]. Journal of Radars, 2017, 6(2): 136–148 doi: 10.12000/JR16130

KANG Miao, LENG Xiangguang, LIN Zhao, et al. A modified faster R-CNN based on CFAR algorithm for SAR ship detection[C]. 2017 International Workshop on Remote Sensing with Intelligent, Shanghai, China, 2017: 1–4. doi: 10.1109/RSIP.2017.7958815.

WANG Chonglei, BI Fukun, ZHANG Weiping, et al. An intensity-space domain CFAR method for ship detection in HR SAR images[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(4): 529–533 doi: 10.1109/LGRS.2017.2654450

LZZO A, LIGUORI M, CLEMENTE C, et al. Multimodel CFAR detection in foliage penetrating SAR images[J]. IEEE Transactions on Aerospace and Electronic Systems, 2017, 53(4): 1769–1780 doi: 10.1109/TAES.2017.2672018

熊开玲, 彭俊杰, 杨晓飞, 等. 基于核密度估计的K-means聚类优化[J]. 计算机技术与发展, 2017, 27(2): 1–5 doi: 10.3969/j.issn.1673-629X.2017.02.001

XIONG Kailing, PENG Junjie, YANG Xiaofei, et al. K-means clustering optimization based on kernel density estimation[J]. Computer Technology and Development, 2017, 27(2): 1–5 doi: 10.3969/j.issn.1673-629X.2017.02.001

冷祥光, 计科峰, 宋海波, 等. 影响星载SAR舰船检测的关键因素[J]. 遥感信息, 2016, 31(1): 3–12 doi: 10.3969/j.issn.1000-3177.2016.01.001

LENG Xiangguang, JI Kefeng, SONG Haibo, et al. Key factors influencing ship detection in spaceborne SAR imagery[J]. Remote Sensing Information, 2016, 31(1): 3–12 doi: 10.3969/j.issn.1000-3177.2016.01.001

NOVAK L M, OWIRKA G J, and NETISHEN C M. Performance of a high-resolution polarimetric SAR automatic target recognition system[J]. Lincoln Laboratory Journal, 1993, 6(1): 11–24.

QIN Xianxiang, ZHOU Shilin, ZOU Huanxin, et al. A CFAR detection algorithm for generalized Gamma distributed background in high-resolution SAR images[J]. IEEE Geoscience and Remote Sensing Letters, 2013, 10(4): 806–810 doi: 10.1109/LGRS.2012.2224317

张颢, 孟祥伟, 刘磊, 等. 改进的基于Parzen窗算法的SAR图像目标检测[J]. 计算机科学, 2015, 42(11A): 151–154

ZHANG Hao, MENG Xiangwei, LIU Lei, et al. Improved parzen window based ship detection algorithm in SAR images[J]. Computer Science, 2015, 42(11A): 151–154

DAI Hui, DU Lan, WANG Yan, et al. A modified CFAR algorithm based on object proposals for ship target detection in SAR images[J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(12): 1925–1929 doi: 10.1109/LGRS.2016.2618604

TIAN Sirui, WANG Chao, and ZHANG Hong. An improved nonparametric CFAR method for ship detection in single polarization synthetic aperetuer radar imagery[C]. IEEE International Geoscience and Remote Sensing Symposium, Beijing, China, 2016: 6637–6640. doi: 10.1109/IGARSS.2016.7730733.

张苗辉, 郭拯危, 刘扬. 基于混合模型的SAR影像海陆分割算法[J]. 光电子•激光, 2017, 28(3): 326–333 doi: 10.16136/j.joel.2017.03.0248

ZHANG Miaohui, GUO Zhengwei, and LIU Yang. Sea-land segmentation algorithm for SAR images based on mixture models[J]. Journal of Optoelectronics•Laser, 2017, 28(3): 326–333 doi: 10.16136/j.joel.2017.03.0248

施引文献

资源附件(0)

访问统计